Combining AI-powered image analysis with image-based risk scoring has demonstrated the ability to predict long-term breast cancer risk, according to a new study from the Mayo Clinic and the University of California, San Francisco. .
WHY IT IMPORTANT
In the report published on Journal of Clinical Oncologyresearchers used Screenpoint Medical’s Transpara Checkpoint with Volpara’s TruDensity software to study how this combination could aid radiologists in predicting the long-term risk of late-stage breast cancer. and terminal breast cancer.
The Mayo Clinic and UCSF collected two-dimensional, full-field digital mammograms performed 2–5.5 years prior to cancer diagnosis from 2,412 women with invasive breast cancer and 4,995 controls. from two US mammography cohorts.
They found that using AI-powered image analysis with an image-based risk classifier would allow radiologists to accurately quantify breast tissue, as the imaging software does reduce reader variability in the assessment of breast density.
According to the announcement of new US research, the TruDensity AI algorithm uses a combination of X-ray physics and machine learning to generate an accurate volumetric measurement of breast composition to eliminate detectable variability. born from human interpretation.
Dr Ralph Highnam, Volpara Health’s director of science and innovation, said: “Breast density is an important factor in breast cancer risk assessment and an objective, volumetric density measurement is very important. important.
“Through the power of AI, we can uncover valuable insights that help clinicians identify individuals at risk for cancer and tailor screening and prevention strategies.” personalized.”
This retrospective study echoes an earlier study published in March European radiology merged screening results from women participating in the Breast Screening program in Norway with AI scores and explored consensus, recall, and cancer detection for different theoretical scenarios of AI and radiologist while reading the screen.
The Norwegian study tested this method – using Transpara’s checkpoint and TruDensity’s image analysis – to evaluate 949 screen-detected breast cancers, 305 distance cancers and 13,646 negative test cases conducted between 2010 and 2018.
The US researchers said in their abstract: “We assessed Breast Imaging Data and Reporting System density, AI malignancy score (1-10) and volume density measurements. “.
“We used conditional logistic regression to estimate the odds ratio, 95% CIs, adjusted for age, and BMI and C to describe the association of AI scores,” they said. with invasive cancer and its contribution to breast density measurement models”.
Although AI scores improved predictions for all cancers in densitometric models and improved discriminant ability for advanced cancer, this approach did not reach statistical significance for densitometry. with interstitial or breast cancer detected between mammograms.
The Norwegian researchers note that 20–30% of screen-detected cancers and gap cancers are classified as missed in informative retrospective review studies, and cite Dual-reading protocols and shortages of radiologists in Europe are driving the demand for AI-based diagnostic tools.
“AI systems have been proposed as a tool to assist or even replace radiologists in the reading process,” they wrote in the full report. They were able to compare the AI’s performance with a radiographer’s dual reading.
“The accuracy of the AI system is comparable to that of a single reader in a standalone dual-reading setting,” they said.
TREND TO BIGGER WOMAN
For years, the fight against breast cancer has been complicated by reading mammograms in women with high breast tissue density. The goal is to promote early detection and help women get care faster – while stopping the rising breast cancer mortality rate worldwide.
Breast cancer isn’t the only area where artificial intelligence can help radiologists, and many fear AI will replace doctors.
Michael Donovan, co-founder and chief medical officer at PreciseDx, a medical IT company that uses AI for treatment, although the immediate and long-term benefits of AI could help diagnose and then deliver treatment. care management decisions. personalized medicine.
“The bottom line is that, through the advantages and efficiency of AI-powered care for physicians, the ultimate arbiter, from a medical diagnostic, ethical and legal perspective, is the physician,” he said. speak Healthcare IT News in January.
“For example, AI alone cannot make a final diagnosis on an X-ray scan or make a definitive diagnosis on a patient’s needle biopsy specimen.”
ON PROFILE
“While we’ve known for decades that breast cancer density and risk are correlated, recent research has really boosted our ability to better understand the effects of breast cancer. density combined with image-based risk for personalized medicine for women,” Dr Nico Karssemeijer, chief scientific officer at Screenpoint Medical and professor at Radboud University, said in the statement.
Andrea Fox is the senior editor of Healthcare IT News.
Email: [email protected]
Healthcare IT News is a publication of HIMSS Media.
